Traffic Matrix Estimation via Simulated Annealing Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for estimating source-to-destination traffic matrices in communication networks are inadequate, particularly for large networks, as they rely on unrealistic assumptions and are not feasible for large traffic volumes, and do not allow for easy specification of arbitrary constraints.
Innovation Solution
A simulated annealing method and system (SATME) that estimates traffic matrices by generating a starting matrix, modifying it randomly, and using a fitness function to converge to a minimum, allowing for the inclusion of arbitrary constraints and providing a probability distribution of traffic patterns.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If packet monitoring and recording methods are used to obtain traffic statistics, then traffic matrix information can be obtained, but huge storage tables are required and it becomes infeasible for large networks and large numbers of packets
Solution Approach 1:
The patent extracts only the necessary traffic count information from packets at network ingress and egress points, rather than monitoring and storing details of every packet. This selective extraction obtains sufficient traffic matrix information while avoiding the need for huge storage tables that would be required for complete packet monitoring.
Solution Approach 2:
The patent segments the traffic measurement problem into two parts: (1) collecting ingress and egress traffic counts at network boundaries, and (2) using mathematical optimization to infer the complete traffic matrix. This segmentation avoids the need to process and store individual packet information while still obtaining accurate traffic matrix data.
2Device complexity
If gravity model or tomographic model are used to infer traffic matrices, then problem size is reduced, but unrealistic assumptions about traffic patterns must be made
Solution Approach 1:
The patent changes the approach from using fixed traffic pattern models (gravity model, tomographic model) to using an optimization-based approach that infers traffic matrices directly from measured link counts. This parameter change eliminates the need for unrealistic traffic pattern assumptions while maintaining computational feasibility through mathematical programming.
3Ease of operation
If link counts are used to derive traffic matrix, then measurement is simplified, but mathematical certainty cannot be achieved and only probabilistic inference is possible
Solution Approach 1:
The patent uses an optimization framework that incorporates feedback from measured link counts to iteratively refine the inferred traffic matrix. The objective function compares inferred link traffic with actual measurements and adjusts the traffic matrix to minimize differences, achieving mathematical certainty rather than just probabilistic inference.
Data Source
AI summary
The SATME method and system estimates source-to-destination traffic matrices using a simulated annealing algorithm, the traffic matrix estimation being represented as a probability distribution over the set of all possible matrices that satisfy a set of given constraints. The constraints explicitly encode information that the user knows about the network traffic as components of an objective function (a fitness function), that is then minimized using simulated annealing. With the method according to the invention, arbitrary constraints of any form can be included, and the case where there are no feasible solutions can be diagnosed by the objective function not converging to zero.